On Thu, 2011-02-17 at 06:52 +0100, Svante Signell wrote:
> On Thu, 2011-02-17 at 01:04 +0100, Samuel Thibault wrote:
> > Diego Nieto Cid, le Wed 16 Feb 2011 14:33:15 -0300, a écrit :
..
> I will integrate your patch with my changes to a create proper patch.
Almost complete by now, only dhclient-sc
Svante Signell, le Tue 22 Feb 2011 00:38:03 +0100, a écrit :
> gcc -g fails_on_hurd.c
> ./a.out
> getifaddrs: (os/kern) successful
It's a bug in pfinet, I'll handle that.
Samuel
Hello,
The getifaddrs bug is due to the following:
/* Return a list of interfaces as expected by the SIOCGIFCONF ioctl.
The maximum number of bytes returned can be given in AMOUNT, but if
AMOUNT is negative, all interfaces will be returned. Always
succeeds (and interface list will be tru
ssize_t is for things that either return a size_t value or -1.
It's never used for parameters. I don't think a vm_ssize_t makes any sense.
This is an odd way to use constructed the RPC interface for this, but we
already have it and perhaps there was some reason to do it that way.
You can just pas
Roland McGrath, le Mon 21 Feb 2011 16:42:23 -0800, a écrit :
> ssize_t is for things that either return a size_t value or -1.
> It's never used for parameters. I don't think a vm_ssize_t makes any sense.
>
> This is an odd way to use constructed the RPC interface for this, but we
> already have i
> So we need to change the RPC documentation to explicit "-1" instead of
> "negative", right?
Well, sure. There is nothing magical about -1, it's just that it will give
the largest maximum that's possible. So you could remove the remark about
a negative value entirely, or write something else th
Svante Signell, le Tue 22 Feb 2011 00:38:03 +0100, a écrit :
> gcc -g fails_on_hurd.c
> ./a.out
> getifaddrs: (os/kern) successful
To fix it you need both the hurd patch I've just pushed and the glibc
patch Roland has just pushed.
Samuel
On Tue, 2011-02-22 at 02:04 +0100, Samuel Thibault wrote:
> Svante Signell, le Tue 22 Feb 2011 00:38:03 +0100, a écrit :
> > gcc -g fails_on_hurd.c
> > ./a.out
> > getifaddrs: (os/kern) successful
>
> To fix it you need both the hurd patch I've just pushed and the glibc
> patch Roland has just pu
